1. Field of the Invention
The present invention relates to a semiconductor memory cell consisting of two transistors or one transistor formed by combining two transistors.
2. Description of the Prior Art
FIG. 17 shows a conventional dynamic memory cell consisting of a single transistor and a capacitor; the memory cell of this structure is generally known as a one-transistor memory cell and used as a semiconductor memory cell suitable for high density devices. In such a memory cell, the capacitance of the capacitor needs to be high enough to cause a voltage change on the associated bit line. However, the trend to smaller plan area of the semiconductor memory cell necessitates a reduction in the size of the capacitor formed in a horizontal plate-like shape, giving rise to the problem that when reading information stored as a charge on the memory cell capacitor, the information is masked by noise, or that, because of increasing stray capacitance on the bit line with each generation of memory cells, only a small voltage change can be caused to the bit line. In one approach to solving this problem, there is proposed a dynamic memory cell having a trench capacitor cell structure (see FIG. 18) or a stacked capacitor cell structure. However, because of processing limitations on how deep the trench can be formed and how high the stack (stacked layers) can be made, there is a limit to increasing the capacitance of the capacitor. Accordingly, it is said that dynamic memory cells of these structures encounter the limit of miniaturization in the dimensional region beyond the low submicron rule.
Furthermore, with transistors forming the semiconductor memory cells, reducing the transistor's plan area beyond the low submicron rule would introduce such problems as dielectric strength degradation, punch-through, etc., which would increase the possibility of leakage even under the rated voltage condition. With conventional transistor structures, therefore, it becomes difficult to ensure proper operation of the memory cell when the memory cell size is reduced.
To overcome the above limitations on the capacitance, there is proposed a memory cell structure wherein the memory cell is constructed with two transistors and the transistor channel current is sensed.
In the memory cell disclosed in Japanese Patent Unexamined Publication No. 63-19847, for example, a capacitor C.sub.2 coupled to the gate and drain of a MOS transistor Q.sub.1 is connected to a word line, as shown in FIG. 3 accompanying the same Patent Publication. Furthermore, the drain of the transistor Q.sub.1 is connected to the gate of an SOI transistor Q.sub.2 of the complementary type to the transistor Q.sub.1. The drain of the transistor Q.sub.2 is in turn connected to a fixed potential V.sub.D, while the source of each of the transistors, Q.sub.1 and Q.sub.2, is connected to a bit line.
In the memory cell disclosed in the above Patent Publication, since the gate electrode of the SOI (silicon-on-insulator) transistor Q.sub.2 is formed only on one principal surface side of the channel region thereof having two principal surfaces, an extra capacitor C.sub.2 is needed to complete the structure of the memory cell. A further problem is that because of the charge or electric field applied (through an insulating film) to the other principal surface of the channel region of the SOI transistor Q.sub.2, the operation of the SOI transistor Q.sub.2 becomes unstable and it is difficult to reduce the channel length.
On the other hand, the memory cell disclosed in Japanese Patent Unexamined Publication No. 1-145850 is constructed with a write transistor 18, a read transistor 19 (an SOI transistor), and a protective capacitor 20, as shown in FIG. 1 accompanying the same Patent Publication. The source of the write transistor 18 is connected to the gate of the read transistor 19. The cathode of the protective diode 20 is connected to the drain of the read transistor 19. Further, the anode of the protective diode 20 and the drain of the write transistor 18 are connected to a bit line, while the source of the read transistor 19 and the gate of the write transistor 18 are connected to a word line.
In the memory cell disclosed in the above Patent Publication, since the read transistor 19 is an SOI transistor, when the source potential is caused to change to the same polarity as the gate potential (when writing a "1") the potential of the channel region changes with the change of the gate potential, resulting in an incomplete writing condition. It is therefore difficult to read the "1" with the expected certainty. Furthermore, in the read transistor 19, the gate electrode is formed only on one principal surface side of the channel region having two principal surfaces. Therefore, this structure also has the problem that because of the charge or electric field applied (through an insulating film) to the other principal surface side of the channel region of the read transistor 19, the operation of the read transistor 19 becomes unstable and it is difficult to reduce the channel length.
The memory cells disclosed in Japanese patent Unexamined Publication Nos. 62-141693 and 62-254462 each comprise a write transistor T1, a read transistor T2, and a storage capacitor C, as shown in FIG. 1 accompanying the former Patent Publication. The drains of the transistors T1 and T2 are connected to a bit line, and the source of the transistor T1 is connected to the storage capacitor and also to a first gate of the transistor T2. Further, the gate of the transistor T1 is connected to a write selection line, while a second gate (or a channel forming region such as a well) of the transistor T2 is connected to a read selection line.
When the read transistor T2 is formed from an SOI transistor, the first and second gates are respectively formed on the upper and lower surface sides of the channel region of the transistor T2 (refer to Japanese Patent Application No. 55-93521 and Japanese Patent Unexamined Publication No. 57-18364). This structure eliminates the problem that has placed a limitation on the reduction of the channel length. For the write transistor T1, on the other hand, only bulk-type transistors are disclosed as preferred embodiments, and therefore, there is a limit to the miniaturization of the memory cell as a whole. Furthermore, the structure requiring each word line to be divided into a read line and a write line has the problem of increased chip area or increased number of stacked layers.
The present invention is concerned with the structure of a memory cell constructed with two transistors and yet capable of solving the above enumerated problems. An object of the invention is to provide a semiconductor memory cell, a semiconductor memory cell for ASICs (Application Specific Integrated Circuits), and even a semiconductor memory cell with a single transistor formed by combining two transistors, of the structure that ensures stable transistor operation, that does not require the provision of a large-capacitance capacitor as required in prior art DRAMs, and that allows the channel length to be reduced and achieves miniaturization of the cell.